Unlike DARPA, ARPA-E does not create technology that the government wants to buy.
This ‘three party system’ creates many problems. There is already an energy market with large entrenched players. This is not so different from a military - people build their careers on specific warfare paradigms and spend a ton of money on equipment that could be obsoleted by new paradigms. So where is the difference? Some possibilities include: DARPA is more connected into the military than ARPA-E is into the energy companies (Manage the Transfer not the Technology) so it’s easier to find champions.
Sometimes DARPA works on programs that go against the established paradigm but sometimes they work on things that the military ‘customer’ wants. Speculatively, this balance between working on weird shit and desired things may make the weird shit more likely to go through? The military, while they are driven by fair prices, are not driven by profit so they don’t make tradeoffs the same way the energy market does. The energy industry won’t make market precommitments for fusion power because they can’t know how to price them because it depends on what the price of oil and gas will be in the future.
The commercialization issues facing ARPA-E need to be considered for any riff on the §ARPA model. If something needs to be integrated into preexisting systems or manufactured, a pipeline into those systems is important!
ARPA-E could get around this in one of two ways: being hyper aggressive in trying to create entirely new industries (decentralized power?) or establishing pipelines into the existing industries. Instead of technology, one of the most powerful things ARPA-E might be able to do is to establish something analogous to FDA trials.